Fine-grained navigation is becoming increasingly demanded in urban areas. However, current navigation software fails to provide accurate paths on grounds without roads (GWR) like residential communities, squares, parks, etc. In this work, a method for the construction of fine-grained navigation networks for urban environments is developed. Remote sensing images are used to estimate the traversability of GWR. A Voronoi diagram that conforms to important shapes, such as buildings, is created for the spatial partitioning of GWR into meaningful atomic regions. A computational model of traversability is constructed based on the Voronoi diagram to construct the navigation mesh of GWR. A fine-grained navigation network is then generated by integrating the GWR navigation mesh and urban road networks. Two experiments are performed to evaluate the effectiveness and efficiency of the proposed method. The results show that the proposed method performs well in both the representation of traversability in GWR and fine-grained path planning in urban areas. Moreover, the integration of the fine navigation mesh and urban road networks proves that such a method has the potential to provide fine-grained navigation services.

城市地区对细粒度导航的需求越来越大。然而，目前的导航软件无法在没有道路 (GWR) 的场地（如住宅社区、广场、公园等）上提供准确的路径。在这项工作中，开发了一种为城市环境构建细粒度导航网络的方法。遥感图像用于估计 GWR 的可穿越性。创建符合重要形状（例如建筑物）的 Voronoi 图，用于将 GWR 空间划分为有意义的原子区域。基于Voronoi图构建遍历性计算模型，构建GWR的导航网格。然后通过集成 GWR 导航网格和城市道路网络来生成细粒度的导航网络。进行了两个实验来评估所提出方法的有效性和效率。结果表明，所提出的方法在 GWR 中的可遍历性表示和城市区域的细粒度路径规划方面表现良好。此外，精细导航网格和城市道路网络的集成证明了这种方法具有提供细粒度导航服务的潜力。